Quick Answer
A short with an avatar takes two inputs and about five minutes. One photo where your face fills at least a third of the frame, and 15 to 30 seconds of clear speech; samples from about 5.6 seconds are accepted, but longer carries more of your voice. The avatar costs 5 credits for the face and 5 for the voice, once, which the free plan's 10 credits cover, then video runs 2 credits a second, so a 45 second short is 90 credits, about $1.90 on the Creator plan. Captions add 1 credit a second, and realistic AI video has to be labeled.
Make a short with an AI avatar: the actual input requirements, the five steps, what each length costs at 2 credits a second, and the 3 minute ceiling on both Shorts and Reels. Checked 15 September 2026.
Related next steps
Nobody can promise you reach, so this page does the part that can be done reliably: turning a photo and a voice sample into a publishable short, with the real specifications and the real cost. Percify numbers were read on 15 September 2026, platform limits on 16 September 2026.
The specifications, stated correctly
- One photo, with your face filling at least a third of the frame. A group shot or a distant photo does not have enough face to animate.
- 15 to 30 seconds of clear speech for the voice. Samples from about 5.6 seconds are accepted, though a longer one carries more of your voice.
- That is the whole input. No studio, no camera, no script recording.
If you have read "30 seconds of your voice" as a requirement on this site before, that was wrong, and this page is the correction.
The steps
- Make the avatar once. Clone Yourself: 5 credits for the face, 5 for the voice. The free plan's 10 credits cover both, so the first one is free.
- Write for the length you will publish. A 45 second short is the useful range; anything under 60 seconds is cheap to iterate on.
- Render. InfiniteTalk Fast ↗ is 2 credits a second: 30 seconds is 60 credits, about $1.27 on Creator, and 45 seconds is 90, about $1.90.
- Caption it at 1 credit a second, because feeds play silent.
- Label it before posting: TikTok requires a label ↗ on realistic AI content, including AI generated speech.
Test the opening without spending
The first line decides whether the rest is watched. Percify's UGC hooks render for no credits, within a daily allowance that grows with the plan, so you can try ten openings and only pay to render the winner as a full short. That is the single cheapest habit in this workflow.
How long a short can be
YouTube's help pages say its Shorts tools make videos "up to 3 minutes long" ↗. Instagram's help pages say Reels over 3 minutes are not recommended to new audiences. So 3 minutes is the ceiling on both, and at 2 credits a second a full 3 minutes would be 360 credits, about $7.59, which is another reason the 30 to 45 second range is where this workflow pays.
What decides reach beyond length is not something anyone outside the platforms can tell you, but TikTok does publish what it uses, including the fact that follower count is not a direct factor: what TikTok says decides reach.
What this workflow does not include
For voices, use your cloned voice or a preset voice from the text to speech models in the playground, such as ElevenLabs v3 and MiniMax Speech; Zonos 2 reads a script at 2 credits a minute. For another language, the playground dubs through ElevenLabs Dubbing at 72 credits a minute, or HeyGen video translation at 3 credits a second into 178 output languages. There is no scheduling and no analytics either; posting happens in the apps.
More: AI avatars for Instagram Reels, each platform's label rule, the best AI avatar generators compared and Percify's plans.
Ready to Create Your Own AI Avatar?
Make a talking video of yourself from one photo, in your own voice. No watermark on any plan.
Cancel anytime from your profile
Got questions?
Frequently asked
Make the avatar once from one photo and 15 to 30 seconds of speech (5 credits for the face, 5 for the voice, covered by the free plan's 10 credits), then render at 2 credits a second, caption at 1 credit a second, and label it before posting.
On Percify, 15 to 30 seconds of clear speech is the recommendation, and samples from about 5.6 seconds are accepted. A longer sample carries more of your voice. Anything claiming a flat "30 seconds" requirement is repeating an error.
90 credits of video, about $1.90 on the Creator plan ($26 for 1,233 credits), plus 45 credits for captions at 1 credit a second. The avatar itself is a one time 10 credits, which the free plan covers. Read on 15 September 2026.
YouTube's help pages say its Shorts creation tools make videos up to 3 minutes long, and Instagram says Reels over 3 minutes are not recommended to new audiences. At 2 credits a second a full 3 minutes is 360 credits, about $7.59.
Yes. The text to speech models in the playground, such as ElevenLabs v3 and MiniMax Speech, offer preset voices, or you can clone your own voice for 5 credits once. Zonos 2 reads a script at 2 credits a minute.
